Some questions regarding svd computations

31 Views Asked by At

0

I've some questions regarding svd:

Consider a matrix A ∈ R500×5 and its SVD [U, S, V T ] = svd(A).

(Assume A is centered).

a) is the second left singular vector of A is the direction in R5 with the second most variance?

b) is the first right singular vector of A is the direction in R5 with the most variance.

Let u1, u2 be the first two left singular vectors; let v1, v2 be the first two right singular vectors; and let s1, s2 be the first two singular values.

Consider $B = s1*u1*u^{T}_1+s2*u2*v^{T}_2$.

c) What is the dimension of B? d) What is the rank of B?